Although intelligent cockpit and intelligent driving have become a big hit in the automotive industry, in the past two years, only a handful of models under 200,000 have high-end intelligent driving capabilities, let alone any user experience. In promoting the popularization and decentralization of intelligent driving, someone urgently needs to stand up first.
At present, the cost of mainstream high-end intelligent driving solutions in the industry is still very expensive, which is the core reason why low-cost models are difficult to support. Based on the analysis of the self-developed solutions of most car companies, they want to achieve L2
At the same time, under the situation of extreme involution in market competition, either car companies will "draw cakes" for users in advance, and the final function push will be far away; Either the capable car companies "reduce the price of packaging", such as the price of 3The Huawei ADS high-end system package of 6W yuan is discounted to 19w yuan, or even 3000 yuan; Xiaopeng is directly attached to the car with a price difference worth 20,000 yuan; NIO will give a one-year or two-year NOP+ experience for a limited time.
In short, at this stage, high-end intelligent driving is not only a long-term core competitiveness that cannot be ignored for car companies, but also a selling point of "losing money and making money", even if it is not possible to recover the cost in the ** model of 20 300,000 yuan, how can it be further delegated to low-cost models?
However, with the further development of technology, intelligent driving cost reduction has gradually become the next mainstream trend in the industry. In particular, the emergence of DJI vehicle-mounted intelligent driving in China is known for its low-cost characteristics, which has given rise to the signs of "decentralization of intelligent driving".
DJI has joined forces with BYD and FAW
Recently, according to 36Kr, DJI has received a clear investment intention from BYD and FAW, and all parties are currently in the financing process, and will jointly develop autonomous driving solutions in the future. According to reports, the valuation of DJI's foreign investment institutions is about 10 billion yuan.
BYD sold a total of 302 in 2023440,000 new energy vehicles, a year-on-year increase of 623%, is the highest sales champion of new energy vehicles in China and even in the world. Entering 2024, BYD clearly released a signal to increase intelligence, and demonstrated a series of core technologies including vehicle intelligence, intelligent driving, intelligent parking, and intelligent cockpit at the dream day in January.
In terms of intelligent driving, BYD has adopted multiple sets of solutions in parallel, including Yangwang's full-stack self-development, the cooperation between DENZA and Momenta, and the introduction of Horizon. In the ** report, it was also mentioned that BYD had intended to contact Huawei Car Bu, but the latter believed that there was not much room for cooperation between the two sides and politely declined.
Although it was rejected by Huawei, BYD found a DJI cooperation this time. DJI Automotive is currently a veritable "sweet spot" in the industry, because it has launched a new generation of intelligent driving solution "Chengxing Platform", which can realize high-end functions such as urban NOA and commuting mode without relying on lidar and high-precision maps, and focuses on the ultimate cost-effective advantages of low cost and high income.
At the 2024 International Consumer Electronics Exhibition, DJI Automotive demonstrated that based on a single Qualcomm SA8650P autonomous driving-grade chip, 12 cameras, 12 ultrasonic radars, and 1 millimeter-wave radar, it can realize high-end intelligent driving solutions at the NOA level in urban areas, and announced that it will be officially mass-produced soon.
Compared with competitors who often carry lidar and hundreds of tons of computing power Orinx chips, DJI only needs a minimum of 7 9 cameras with 32T computing power to achieve urban memory driving (commuting mode); Upgrading the chip to 80T computing power can support NOA in urban areas, which shows that DJI Intelligent Driving has squeezed the hardware cost to the minimum.
Of course, no matter how much you say, there is no real car experience that is more shocking. It is the first to carry the DJI Lingxi Intelligent Drive 20 Baojun Cloud Lingxi Edition has been on the market for many months now, and the car is priced at less than 1340,000 yuan, but it can provide national mapless high-speed piloting, memory parking within 1km, tracking and reversing within 100 meters, L2+ automatic assisted driving within the range of 0-130km H, etc., both in terms of functional richness and actual user experience, are more powerful than previous models at the same price.
Behind the low cost is the difference in technology
Some people may ask, now that the whole industry is vigorously studying intelligent driving, why is only DJI able to make a cost-effective solution?
The answer is that there are differences in technical routes. At the most critical perception level of autonomous driving, the DJI Chengxing platform adopts the industry's rare inertial stereo binocular front-view camera module, draws on the principle of human binocular imaging, and "sees" a deep three-dimensional space through the principle of binocular fusion parallax, thus directly replacing the role of lidar.
It has extremely strict accuracy requirements for almost all elements such as the relative position, angle, image stabilization, and structural change after long-term operation, visual algorithm calibration, etc., which may lead to image recognition errors even if it is a millimeter, which is also the reason why few manufacturers in the industry have adopted it.
However, for DJI, this is the most familiar and convenient thing, because the inertial navigation stereo binocular has long been widely used in the field of DJI drones, and it is also the technical foundation for DJI drones to dominate the world.
In addition, DJI can also use stereo binocular to perceive the road environment information ahead in real time, and provide necessary information for the vehicle's CDC electromagnetic suspension, adaptive headlights, windshield wipers, etc. Whatever LiDAR can do, it can do too.
Finally, coupled with DJI's on-board BEV deep neural network algorithm and fully self-developed open space decision-making technology, it not only gets rid of the heavy dependence on high-precision maps, achieves real-time local mapping, but also greatly reduces the real-time computing load of the chip, eliminating the need for high-computing chips, and further reducing costs.
There is no doubt that DJI's vehicle-mounted Chengxing platform has significant advantages such as low cost, strong functions, and good experience, and with it, even the Baojun Yunduo Lingxi Edition in the early 100,000 yuan can have the same high-end intelligent driving experience as the mid-to-high-end models of 200,000 yuan. Next, if it can successfully join forces with BYD and FAW, then the new cars below 200,000 yuan will be equipped with high-end intelligent driving, which will no longer be an empty dream, and the equal right of intelligent driving will also really accelerate the popularization.